There are a number of applications for techniques for optical measurement through light scattering materials. Most notably, such measurements can be performed through biological tissues and therefore can be used for noninvasive medical diagnostic tests. Cancer tissue and healthy tissue, for example, can be distinguished by means of different optical properties. Scanning the optical measurement can yield high contrast and high magnification images of biological tissues. For example, imaging techniques could be used to examine plaque on the interior walls of arteries and vessels or other small biological structures. Related applications extend to the examination and troubleshooting of integrated optical circuits, fiber optic devices, and semiconductor structures. All these applications require that the measuring technique have a relatively high spatial resolution (microns or tens of microns), high sensitivity, and low noise.
Optical time domain reflectometry (OTDR) and optical frequency domain reflectometry (OFDR) are techniques which are used to examine optical systems and are generally not capable of performing high resolution measurements through a light scattering material. For example, these methods are generally designed for finding and locating (to within 1 meter) flaws in a fiber optic system.
Optical coherence domain reflectometry (OCDR) is a technique which has been used to image an object within or behind light scattering media. The technique uses short coherence length light (typically with a coherence length of about 10-100 microns) to illuminate the object. Light reflected from a region of interest within the object is combined with a coherent reference beam. Interference occurs between the two beams only when the reference beam and reflected beam have traveled the same distance. This allows the OCDR to discriminate against light scattered from outside the region of interest.
FIG. 1 shows a typical OCDR setup similar to ones disclosed in several U.S. patents (U.S. Pat. Nos. 5,465,147, 5,459,570, and 5,321,501 issued to Swanson et al., U.S. Pat. Nos. 5,291,267, 5,365,335, and 5,202,745 issued to Sorin et al). FIG. 1 shows the device made with fiber optic components, but OCDR devices can also be made with bulk optical components. Light having a short coherence length l.sub.c (given by l.sub.c =C/.DELTA.f, where .DELTA.f is the spectral bandwidth) is produced by a light source 20 and travels through a 50/50 coupler 22 where it is divided into two paths.
One path goes to the sample 24 to be analyzed and the other path goes to a movable reference mirror 26. Extra fiber length in the reference path is shown as fiber loop 31. The probe beam reflected from the sample 24 and reference beam reflected from the reference mirror 26 are combined at the coupler 22 and sent to a detector 28. The optical paths traversed by the reflected probe beam and reference beam are matched to within one coherence length such that coherent interference can occur upon recombination at the coupler.
A phase modulator 30 (such as a piezoelectric fiber stretcher) produces sideband frequencies in the probe beam which produce a temporal interference pattern (beats) when recombined with the reference beam. The detector 28 measures the amplitude of the beats. The amplitude of the detected interference signal is a measure of the amount of light scattered from within a coherence gate interval 32 inside the sample 24 that provides equal path lengths for the probe and reference beams. Interference is produced only for light scattered from the sample 24 which has traveled the same distance (to within approximately one coherence length) as light reflected from the mirror 26. The coherence gate interval 32 has a width of approximately one coherence length. This feature of OCDR allows the apparatus to discriminate against light which is scattered from outside the coherence gate interval 32, and which is usually incoherent compared to the reference beam. This discrimination (a `coherence gate`) results in improved sensitivity of the device.
One negative consequence of the geometry of FIG. 1 is that 50% of the light reflected from the sample 24 is lost. On its return trip through the coupler 22, half the reflected probe beam enters the light source 20 and does not enter the detector 28. This is undesired because it decreases the signal to noise ratio of the device and results in a more powerful light source being required. Another negative feature of the device of FIG. 1 is that it requires the use of a moving mirror to scan longitudinally in and out of the sample 24. The use of a moving mechanical mirror is a disadvantage because moving mechanical parts often have alignment and reliability problems.
Another disadvantage of the device of FIG. 1 is the requirement for a large depth of focus of the probe beam in sample 24. A large depth of focus is necessary to allow longitudinal scanning of the coherence gate interval 32 while maintaining the coherence gate interval in the region of the beam having a reasonably small spot size. This requirement increases the minimum spot size of the beam, and thus limits the spatial resolution of the device when acquiring images.
A further disadvantage of the device of FIG. 1 is the long integration time typically necessary for each measurement point (pixel) when acquiring an image. This is due to the low power of the backreflected signal when imaging deep within a scattering medium. Under these conditions, the slow acquisition time does not allow in-vivo imaging of live tissue which is usually in motion.
U.S. Pat. No. 5,291,267 to Sorin et al. discloses a technique for OCDR which uses the light source as a light amplifier in order to boost the reflected signal from the sample. Light reflected from the sample is returned through the light source in a reverse direction and is amplified as it passes through. However, Sorin's device requires a coupler in the light path between the source and sample and so necessarily wastes 50% of the light reflected from the sample. In other words, only 50% of the light reflected by the sample is amplified and contributes to the interference signal. Consequently, Sorin's device produces less than optimum signal to noise ratio resulting in less accurate measurements.